lkml.org 
[lkml]   [2004]   [Oct]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    SubjectRe: Totally broken PCI PM calls
    From
    Date
    Hi.

    On Tue, 2004-10-12 at 07:37, David Brownell wrote:
    > On Monday 11 October 2004 2:17 pm, Nigel Cunningham wrote:
    > > On Tue, 2004-10-12 at 02:36, David Brownell wrote:
    > > > I've made that point too. STD is logically a few steps: quiesce system,
    > > > write image to swap, change power state.
    >
    > I'm hoping you agree with that abbreviated summary of
    > what's involved! Pavel seemed to. Of course the devil is
    > in the details, which I hope to leave mostly to others ... ;)

    It certainly fits for Pavel, but, I do have things slightly different,
    so that don't have the maximum-image-size-is-half-of-RAM problem. At the
    moment I do:

    quiesce system & prepare metadata
    suspend devices not used for writing the image or user I/O (this is just
    seeking to remove any chance of them allocating memory, which might
    strangle writing the image)
    write LRU pages ('pageset 2')
    suspend used devices
    save CPU context & make atomic copy of remaining pages
    resume used devices
    save image
    power down used devices
    power down system/enter S4.

    This is why I made that 'device tree' patch - so I could separate the
    devices used for writing the image from those unused and treat them
    separately. I'm pretty sure that I could get away with leaving the
    unused ones alone until snapshot time, but it seems more ideal to me to
    get them to save state and power down at the start, especially if you're
    trying to suspend when the battery is low!

    At the very least, I'd like to see that 'snapshot' state implemented
    separately to the 'powerdown' state.

    > Of course the ACPI spec muddies the water by talking about two
    > different states called "S4": "S4 Sleeping", which is what I was
    > talking about as G1/S4; and "S4 Non-Volatile Sleep" that's more
    > what I've seen with swusp: more like a G2 or G3 poweroff.

    Okay. I've only looked at the ACPI spec occasionally, and have generally
    just followed the lead of Patrick and Pavel in implementing ACPI support
    in s-t-d.

    > I'm willing to believe that there are systems on which swsusp
    > tells drivers a less confusing story ... but I don't seem to have
    > tested with those!

    :> I'm confused by all these changes; no wonder drivers are!

    Nigel
    --
    Nigel Cunningham
    Pastoral Worker
    Christian Reformed Church of Tuggeranong
    PO Box 1004, Tuggeranong, ACT 2901

    Many today claim to be tolerant. True tolerance, however, can cope with others
    being intolerant.

    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

    \
     
     \ /
      Last update: 2005-03-22 14:06    [W:3.103 / U:0.180 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site